Up to four models
DeepSeek-V3 vs Llama 4 Maverick
On the index (GPT-4 = 100), DeepSeek-V3 scores 119 and Llama 4 Maverick 116. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
119103–135
WeightsOpen
ReleasedDec 2024
Price in / out$0.26 / $1.03
Context164K
Core tests taken2 of 9
116100–132
WeightsOpen
ReleasedApr 2025
Price in / out$0.35 / $1.0
Context524K
Core tests taken3 of 9
DeepSeek-V3 and Llama 4 Maverick are too close to call (3 points apart, within the uncertainty), ahead on 0 of 1 shared core tests, at a similar price.
Test by testthird-party results
Every result as a table
| Test | DeepSeek-V3 | Llama 4 Maverick |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 56.5% | 67.0% |
| SimpleBench | 18.9% | 27.7% |
| Maths | ||
| AIME-style maths (OTIS mock) | 15.8% | 20.6% |
| MATH (level 5) | 64.8% | 73.0% |
| FrontierMath (2025 set) | 1.7% | 0.7% |
| Coding | ||
| Aider Polyglot | 48.4% | 15.6% |
Task length · METR
28 sec
Tests at least two of these models have taken. Bold rows are the core tests.